“…Given that information may be available in different spatial scales, one way in which humans may outperform machine identification algorithms is by combining information across spatial scales. Existing computer models tend to rely on level 2 (classical minutiae) features (Champod & Margot, , ; Egli, Champod, & Margot, ; Fang, Srihari, Srinivasan, & Phatak, ; Neumann et al., , ; Srihari & Su, ; Su & Srihari, , , ), though some more recent proprietary AFIS algorithms may use more pixel‐based information.…”